You've built a game that requires a computerised opponent, but aren't sure how to move beyond basic rules-based systems.
How can you build a smart AI that learns the optimal strategy for a game, by repeated play and reinforcement learning?
Our in-house specialists in game AI can help you to design and build a system that is able to learn progressively more complex and sophisticated strategies for your game.
We utilise the cutting edge research from Google DeepMind to build AIs that utilise deep reinforcement learning and Q-learning to evaluate the value of a given action from the board position and game state.
We pit the AIs against eachother, with each successive generation finding new and more elaborate tactics to win the game.
Code that implements the game mechanics An interface to your game
1. A set of game AIs at varying ability levels, coded in Python and Tensorflow, that accept a game state and output the chosen move, along with statistics to explain the relative value of the possible actions.
2. A REST API that exposes the decision making code, that could be built into existing applications.
3. Deployment of the REST API into a cloud architecture or local environment.
You game will be enhanced by an AI that uses the latest advances in deep reinforcement learning to provide a worthy opponent to your users.